How fast is fast compared to D's standard library?

fast can be up to 20x faster for tasks like JSON parsing and string splitting by avoiding GC and using SSE instructions. However, it sacrifices validation and safety features present in std.json or std.string.

How to parse JSON with fast in D?

Import fast.json and use json.coordinates.read! with your struct type, as shown in the README example for reading Point3D[] directly from a JSON file without intermediate allocations.

Does fast work with the LDC compiler?

Yes, fast supports LDC, GDC, and DMD compilers, and it's tested with D front-end versions from 2.068 to 2.079, but compatibility with newer versions might be limited.

Is fast safe to use with untrusted input?

No, fast skips input validation for performance gains, so it's not recommended for security-critical applications without additional sanitization layers.

Can I use fast for general string manipulation beyond splitting?

fast primarily optimizes string splitting with SSE; for other operations like concatenation or searching, you'll likely need D's standard library or other specialized libraries.

How to benchmark fast in my own project?

Run the included benchmark via dub with flags like --config=benchmark --build=release --compiler=gdc, as detailed in the README, to measure performance against your use cases.

Open-Awesome

fast.json

A D library providing the fastest possible implementations of common routines like JSON parsing, string splitting, and C interop.

GitHub

111 stars6 forks0 contributors

What is fast.json?

fast is a D library that provides highly optimized implementations of common routines such as JSON parsing, string splitting, and C interop. It focuses on achieving the fastest possible performance by avoiding garbage collector allocations, skipping input validation, and using SSE instructions and stack allocations. The library can deliver up to a 20-fold speed increase in some scenarios compared to standard implementations.

Target Audience

D developers working on performance-critical applications, systems programming, or low-level tasks where maximum throughput and minimal overhead are essential.

Value Proposition

Developers choose fast for its uncompromising focus on speed, offering GC-free operations, SSE acceleration, and direct memory access that outperform standard library routines. It provides a lightweight, high-performance alternative for specific tasks without the bloat of validation or allocation overhead.

Overview

A library for D that aims to provide the fastest possible implementation of some every day routines.

Use Cases

Best For

Parsing large JSON files into D structs with minimal latency
Splitting strings in performance-sensitive path processing
Interfacing with Windows or Linux C APIs efficiently
Writing high-throughput data processing pipelines in D
Benchmarking and optimizing critical code sections
Systems programming where GC pauses are unacceptable

Not Ideal For

Projects requiring robust input validation and error handling for untrusted data
Teams relying heavily on D's garbage collector for automatic memory management simplicity
Applications prioritizing memory safety and security over raw speed
Developers needing a general-purpose library with broad functionality beyond optimized routines

Pros & Cons

Pros

Blazing Fast JSON Parsing

Directly reads JSON files into user-defined structs without intermediate allocations, reducing overhead and achieving up to 20x speed improvements in benchmarks.

SSE-Optimized String Splitting

Uses SSE3 instructions to split strings around delimiters with minimal overhead, ideal for high-throughput path processing tasks.

Efficient C Interop Utilities

Provides charPtr and wcharPtr helpers for seamless and fast calling of Windows and Linux C APIs, avoiding GC allocations.

GC-Free Performance Predictability

Functions avoid garbage collector allocations entirely, ensuring consistent performance without GC pauses for critical code sections.

Cons

Lacks Input Validation

Skips validation to boost speed, which can lead to crashes or security vulnerabilities when processing malformed or untrusted inputs.

Narrow, Targeted Scope

Focuses only on a few optimized routines like JSON parsing and string splitting, so it's not a drop-in replacement for D's standard library.

Potential Unsafe Low-Level Operations

Relies on direct memory access and stack allocations, increasing the risk of bugs such as buffer overflows if used carelessly.

Frequently Asked Questions

Related Projects

dom.d

This is a collection of modules that I've released over the years. Most of them stand alone, or have just one or two dependencies in here, so you don't have to download this whole repo.

Stars570

Forks130

Last commit3 days ago

cerealed

Powerful binary serialisation library for D

Stars94

Forks3

Last commit5 months ago

orange

A serialization library for the D programming language.

Stars72

Forks15

Last commit6 years ago

dproto

D Protocol Buffer mixins to create structures at compile time

Stars38

Forks16